An Approach To Computing Electrical Parameters In The Realtime-supervisory System Of Power Network

The real-time supervisory to such parameters of pivotal nodes in power network as voltage, active power, reactive power and phase angle is of great significance to acquire the static and dynamic characters actually and ensure power network run safely and economically.In this paper, synchronous card PCI2008AD is used to complete the synchronous sampling of multi-signal of voltage and current. Programming with interruption makes the collection of data and the process data self-governed, giving the process of data enough time.Three methods of orthogonal resolution, qusi-synchronous and three-point orthogonal resolution are analyzed in this paper. For orthogonal resolution, the orthogonal components under different harmonics are acquired from plural voltage and current signals by the means of FFT, with known frequency and synchronous sampling. For qusi-synchronous measurement, multi-recursion is used to ensure the precision close to the ideal one of synchronous sampling, but it takes longer time. The computational time is relatively little in the three-point orthogonal resolution method. It has nothing to do with frequency, but not suitable for harmonics. To save time and improve the validity of parameters, the first method is adopted in this paper.To reduce the affect of the frequency-deviate, the orthogonal resolution method based on frequency-estimation is presented in this paper. After estimating the frequency by zero-crossing method, the synchronous sampling data under the frequency of 50Hz are converted into those with the actual frequency by the means of linear interpolation. The experiments show that it can meet the computational precision and take little time.Comparing the methods of zero-crossing, three-point and four-point, zero-crossing method is adopted to estimate the frequency with harmonic and yawp in actual signals. Because zero-crossing method is relatively sensitive to the yawp nearby the zero-point, FIR low-pass filter of eight ranks is used to treat the signals. The experiments show that the precision estimating the frequency with the data after filtering satisfies the technical demands.